Insults to the DNA of living cells can induce diverse lesions, the more lethal being DNA strand breaks and DNA interstrand crosslinks (1). These lesions have been characterized by methods such as alkaline sucrose gradients or alkaline filter elution (2) and, more recently, the single cell gel electrophoresis (Comet) assay (3). DNA damage can initiate a host of cellular responses, among them changes in replication [i.e., bromodeoxyuridine (BrdU) incorporation into DNA (4)], and in damage-signaling proteins such as the double-strand break indicator γ-H2AX (5). These responses can be monitored using flow cytometry or Western blotting, respectively, which are both time-consuming and/or expensive procedures.

Previously, Norwood et al. (6) described a method for measuring telomere length by loading intact cells on slot blots. We adapted this procedure so that a single slot blot can be used to assess not only total genomic DNA damage, but also levels of either BrdU incorporation or γ-H2AX. HCT116 cells were grown for 24 h in 0.06 uCi/mL ¹⁴C-thymidine (Moravek, Brea, CA, USA), followed by treatments to induce DNA damage. When BrdU incorporation was measured, 100 µM BrdU (Sigma Aldrich, St. Louis, MO, USA) was added to cells 0.5 h prior to harvest. Treated cells were removed from plates with trypsin, washed once with phosphate buffered saline (PBS), and adjusted to 1 × 10⁵ cells per mL in PBS. Slot blots were prepared similarly to the method of Norwood et al., except that Hybond N+ (GE Healthcare Bio-Sciences Corp., Piscataway, NJ, USA) was used instead of Zeta probe. Slots were drained under vacuum at a rate of 0.5 mL/min. Briefly, slots were washed with 200 µL of neutralization solution (1.5 M NaCl, 0.5 M Tris, pH 7.5), and 200 µL of cells (2 × 10⁴) were applied to each of three slots per sample and drained. Two hundred microliters of denaturing lysis solution (1.5 M NaCl, 0.5 N NaOH) was applied per slot. After 7 min of lysis, the slots were drained, and 200 µL of neutralization solution were added for an additional 7 min. These slots were used to calculate ¹⁴C-DNA bound to the membrane. During the second 7-min incubation, 200 µL per sample were loaded on each of three additional slots prewetted with neutralization solution. Denaturing lysis solution was not added to these slots which represent total ¹⁴C-DNA. All wells were drained, and DNA was crosslinked to the wet membrane with a UV DNA crosslinker (Stratagene Corp, La Jolla, CA, USA). The wet blots were blocked for 1 h with 5% milk in Trisbuffered saline, 0.05% Tween 20, pH 7.5 (TBS-T) and incubated overnight at 4°C with primary antibodies to either BrdU (1 µg/mL; DAKO, Carpinteria, CA, USA) or γ-H2AX (2 µg/mL; Upstate, Lake Placid, NY, USA), diluted in blocking solution. Blots were washed in TBS-T, then incubated for 1 h in secondary horseradish peroxidase (HRP)–labeled goat anti-mouse antibody (Sigma Aldrich) diluted 1:20,000 in TBS-T. Chemiluminescence was detected on film using Western Chemiluminescent Reagent Plus (PerkinElmer, Waltham, MA, USA). Since the chemiluminescent signal was more than 10 times greater when cells were lysed in denaturing solution rather than neutralization solution (data not shown), denatured samples were used for quantitating immunoblots. γ-H2AX was expressed as a fraction of the control intensity; that is, average sample intensity was divided by average control intensity:

For ¹⁴C detection, the blots were dried and exposed to a Molecular Dynamics phosphorimager screen (GE Healthcare Bio-Sciences Corp.) for 2 h. The ¹⁴C-DNA bound to the three slots was averaged and compared with the average total ¹⁴C to determine the fraction of ¹⁴C-DNA eluted. The fraction of ¹⁴C-DNA eluted was calculated as follows:

Detection of ¹⁴C DNA was linear with 1 × 10⁴ to 5 × 10⁴ cells per well, while optimal chemiluminescence was observed with 2 × 10⁴ cells (data not shown).

Both strand breaks and DNA crosslinks can be detected using this method; this is shown in Figure 1. Figure 1A is a phosphorimage of the ¹⁴C profile after cells treated with increasing concentrations of the DNA double-strand scission enediyne C-1027 (7) were loaded on a slot blot and the fraction of ¹⁴C-DNA eluted was determined. Figure 1B shows that the fraction of ¹⁴C-DNA eluted increased with increasing C-1027 (i.e., increasing strand breaks). The presence of crosslinks should decrease elution of DNA from the filter. Figure 1C shows reduced ¹⁴C-DNA elution when cells were incubated 3 h with 10–100 nM of the DNA interstrand crosslinker bizelesin (8) prior to 50 Gy irradiation (IR) to induce strand breaks, compared with IR alone. Thus, bizelesin crosslinking blocked IR-induced DNA elution from the filter.

Figure 1. DNA strand breaks and DNA crosslinks in HCT116 cells. (Click to enlarge)

While damage to DNA limited its binding to the membrane, uniform Ponceau staining of all samples on the slot blot (data not shown) indicated that total protein binding to the membrane was independent of DNA damage or elution. Figure 2 shows that either replication or levels of γ-H2AX can be assayed on the same blot used to detect DNA damage. Densitometric scans of representative films of BrdU or γ-H2AX detection in cells treated with C-1027 are shown in Figure 2A or 2C, respectively. Figure 2B shows that BrdU incorporation decreased progressively after a 1 h treatment with 0.5–25 nM C-1027, a known replication inhibitor (9). This decrease correlates with an increase in γ-H2AX (double-strand breaks) after a 1 h treatment with 0.2–10 nM C-1027 [see Figure 2, C and D, in agreement with our earlier studies (10)]. The positively charged nylon membrane, Hybond N+ should bind both double and single-strand DNA over a wide size range. That the slot blot can discriminate between damaged and undamaged DNA may reflect earlier reports that membranes may not bind small DNA fragments efficiently (11,12). It also is possible that draining DNA through the slots under vacuum may not allow sufficient time for smaller DNA fragments to bind to the membrane. The slot blot method is not a replacement for more rigorous measures of DNA damage and cellular responses, but is useful as an extraction-free first step in screening agents for effects on these parameters.

Figure 2. Immunoblots of cells treated with C-1027. (Click to enlarge)